Investigation into the
Effectiveness of Filters for use by Intravenous Drug Users
Jennifer Scott, Christine Bond*,
Emily Kennedy and Arthur Winfield
The Robert Gordon University, School of Pharmacy, Aberdeen
*University of
Aberdeen, Department of General Practice
1997
ABSTRACT: Injecting drug users suffer serious health problems due to intravenous administration of insoluble particles from street drugs and tablets. Makeshift filters are used to try to remove particles from the injections. The decision to distribute filters from harm reduction centres cannot be based on evidence of efficacy and safety as no such published data exists. The work presented here is part of a larger study which anticipates to provide such evidence. Laboratory methods were developed based on information gathered from drug users. Injections were prepared using drugs of abuse and filtered either through makeshift or commercially available filters. The resulting solutions were assayed for particle size and number (using Coulter Counter®). Comparisons were made with unfiltered injections. Work is ongoing, preliminary results will be presented and the path of future work highlighted.
Introduction
The aim of this study is to provide evidence from the laboratory of the effectiveness and predicted safety of filters for use by injecting drug users (IDUs). If such evidence can be produced it's use will be two fold: it can support bids for funding from needle exchanges to enable them to distribute filters and it can be used to create harm reduction information for IDUs to advise on effective filtering.
There are many medical complications of intravenous drug use associated with the insoluble particulate content of injections. Examples of these include: phlebitis, sterile abscesses, blocked blood vessels (which can lead to deep vein thrombosis, varicose ulcers and gangrene) and endocarditis. The latter is more commonly due to bacterial seeding from infected wounds but can be caused by the build up of insoluble adulterants and diluents behind the valves in the heart leading to irritation.
Both the size of the particles and the number present in injections are important. In the microcirculation of the body the smallest vessels are the capillaries, which are approximately 8 microns in diameter. Next are the terminal arterioles which are between 20 and 50 microns. The more particles there are and the bigger they are, the greater the risk to health. Consequently, there are limits placed on the particulate content of commercially made injectables. The British Pharmacoepia stipulates that for large volume parenterals (fluids for intravenous administration, over 100ml), there should be no more than 1000 particles per ml over 2 microns and of these no more than 100 per m should be over 5 microns.(1) There is no limit given in the BY for small volume parenterals. Obviously, the injections prepared by IDUs will contain greater numbers of particles than those made commercially because of the presence of insoluble adulterants in street drugs and diluents in pharmaceuticals and because the IDU is not preparing injections in an aseptic environment. For an IDU's filter to be effective it must reduce the particulate content of the injection to a level that presents a lower risk of harm. It must be acceptable to the user, so must not remove the drug and has to be quick and easy to use. The data presented here are results from the initial part of this study, which investigated the effectiveness of various filters on reducing particulate count and size range of injections made from tablets. The ongoing work is looking at the effect of the filtration methods on amount of drug in injections. It is hoped that work will be able to be done using street heroin, liaisons are currently underway with the Home Office and the police to arrange provision of samples.
Method
Semi structured interviews were carried out with twenty current or ex injecting drug users to establish the drugs that are used by injection and the methods used to prepare injections. From this, the most commonly used drugs by the IV route were established and it became clear that the method of preparation used for heroin was similar, but the preparation method used for tablets varied. Safer Injecting Guidelines published for IDUs were therefore also consulted to assist in establishing the preparation method (2), (3). The Scottish Drug Misuse Statistics (4) were also consulted for information about the most popular injected drugs on a national scale. For financial reasons and constraints on time, it was decided to investigate three pharmaceuticals and one street drug, the most popular drugs being selected for investigation.
The type and quantity of tablets used were as follows: Physeptone ® (methadone 5mg, Glaxo Wellcome) nine tablets in 5mls of water, Diconal ® (dipipanone 10mg and cyclizine 30mg, Glaxo Wellcome) two tablets in 2mls of water and Temgesic ® (buprenorphine 0.2mg, Reckitt and Coleman) five tablets in 2mls of water. The injections were prepared by crushing the tablets between paper with a spoon and mixing the resulting powder on the spoon with water from the kettle which had been boiled and cooled. The methods of filtration investigated were as follows: a piece of a cigarette filter (Lambert and Butler®) prepared by removing the surrounding paper and tearing it in two down the middle then halving one piece with scissors and smoothing any stray fibres with the fingers before placing it in the edge of the solution. A hand rolling filter (Rizla Extras® 7mm acetate filter tips), used whole, placed in the solution. A cotton bud tip (Unichem 100% cotton), removed from the plastic stalk by pulling and the end twisted to smooth the stray fibres then placed in the edge of the solution. In all of theses cases the needle was attached to the syringe and the solution drawn up through the filter. A commercially available syringe filter was also tested (5 micron Acrodisk®, Gelman Sciences). This was placed on the end of the syringe and the solution drawn through, then removed and the needle attached. Also investigated were unfiltered solutions prepared in this way and unfiltered solutions prepared by breaking the tablets in half and shaking them with water in the syringe barrel. Controls were tested using boiled and cooled water and Water For Injections BP (Antigen Pharmaceuticals® ) to investigate the contribution of the filtration and preparation methods on particulate content.
Particle count and size range were measured using the Coulter Multisizer®, using 0.5ml samples withdrawn from the injections immediately after preparation. Each test was repeated three times and the average result taken. It was decided to do this instead of taking three consecutive readings from each sample, as there could be a reduction in the particle count as the sample is stirred. The Multisizer operates on the electrical zone sensing principle: there are two electrodes in a conducting fluid (saline which has been filtered twice using a 0.2micron filter) separated by an orifice of known diameter. A stirrer keeps any particles suspended in the saline.. When a particle passes through the orifice there is a change in resistance (measured as a voltage pulse at a fixed current), the magnitude of this relates to particle size. Total number of particles that are drawn through the orifice in a given time (12 seconds) and number of particles within a size channel can be measured. In this case the 100 micron orifice was used, which measures particles in the size range 2 to 60 microns.
Results
The total number of particles can be determined but is of little value because there will be a large variation in numbers depending on volume of water used to make injections, how finely the drug is crushed etc. Particle size analysis was performed on 0.5m1 sample from the injection. Changes in trends seen in particle size and number before and after filtration were looked for. Before filtration can take place, the drug has to be mixed with water on the spoon. Therefore the number of particles in the unfiltered sample prepared in this way was taken to be 100% and the effects of filtration shown as the percentage reduction in the total number of particles. This is shown in table 1 below. The Physeptone solutions were thick and did not pass through the filters quickly or easily. Several filters had to be used as they clogged and it took longer than was deemed acceptable to the IDU. Since this does not fit with the definition of an effective filter given earlier, it was decided not to continue with the Physeptone filtration investigations.
The total number of particles in the sample is taken to be 100% and the distribution of these according to size range is given for each drug in the tables below:
Discussion
From Table 1 it can be seen than all methods of filtration cause a reduction in the total number of particles in the prepared injections. Preparation by splitting the tablets in two and shaking them in the syringe barrel also reduces the number of particles. However, the concentration of methadone was found to be only 90% of the drug contained in the 9 Physeptone tablets compared to 100% when they were crushed. A slurry was seen to collect at the bottom of the syringe indicating clumping of excipients and drug, explaining why not all the drug dissolved. This material could block the needle or break down enough to pass down the needle and enter the veins as a solid mass. As expected the commercial Acrodisk gave the greatest reduction in the total number of particles. Of the makeshift filters, the Rizla gives the best overall performance for both drugs. The cotton bud showed a good reduction in particle count with Temgesic injections, but was not as effective with Diconal. The cotton bud also has loose fibres which could enter the injections after filtering. Because the Multisizer sizes particles as spheres, a long thin fibre passing through the orifice may be counted as several smaller spheres.
Tables 2a, 2b and 2c show that all filters cause a shift in size range to the smaller end of the scale for the drugs tested. Again, the Acrodisk shows the greatest reduction in size range with the majority of particles in the filtered solution being smaller than the capillaries in the body. Of the makeshift filters, the reduction in size range appears to be greatest with the cotton bud filter. Splitting and shaking the tablets in the syringe also gives more particles in the smaller size ranges. Thus if no filter was used, this would be preferable to crushing the tablets to a fine powder first.
Table 3 shows the total number of particles that were counted without the drug present in the injections as a percentage of the total number detected when the drug is present. This indicates the contribution to the total number of particles that is made by factors such as the kettle, cup, water, spoon, filters, injecting equipment and the non aseptic environment, including the person preparing the injections (e.g. skin shedding). The contribution to particle count not from the drug is low except in the case of kettle water and the Acrodisk. This is because the total number of particles present is low, so the contribution from the equipment and the environment become significant. It is also shown that in terms of particle count, there is no benefit from using Water For Injections. However, there are obvious microbiological advantages.
It has been shown that makeshift and commercial filters reduce the particle count and particle size in injections of Diconal and Temgesic. However, many users do not use filters because they claim to loose the `hit'. Therefore the next stage in this work will develop methods to measure the effects on concentration of filtering. Also, since many more people use heroin than tablets by injection, it is important to repeat this work using samples of street heroin.
References:
2. `A Safer Injecting Guide' HIT, Liverpool, 1995.
3. `What Works?' Exeter Drugs Project, 2nd ed.
4. Information and Statistics Division, The National Health Service in Scotland. Drug Misuse Statistics Scotland (1995 Bulletin). Edinburgh, ISD Publications, March 1996.
Acknowledgements
Sincere thanks are extended to the staff of drugs Action, Aberdeen for their ongoing assistance and support for this project and also to the users for lending their time and knowledge.